Control the grain
Prevent double counting and verify every join before calculating a KPI.
Available learning pathIntermediate
Analysis · KPIs · Quality
Hands-on SQL training for Data Analysts
Learn to turn a business request into reliable SQL analysis: control the grain, build KPIs, use window functions and explain results clearly.
Prerequisites : Master SELECT, filters, aggregates, joins and CTEs.
10structured lessons
10canonical experiences
2progressive modules
5 h 19Estimated duration
Target skills
Essential SQL skills for a Data Analyst
Build KPIs
Produce cohorts, funnels, RFM segments and reproducible time indicators.
Diagnose quality
Detect duplicates, missing values and breaks that distort an analysis.
Deliver a conclusion
Present the result, its checks and limitations in clear business language.
Evidence of progress
A rich learning path is not measured by page count alone.
You learn a method, reuse it on different data, then demonstrate independence in an assessment and a business case.
Understand
Define grain, KPI, denominator and control before calculating.
10 structured lessonsPractise
Solve cases involving windows, cohorts, quality, margin and double counting.
10 canonical experiences · Progressive hints, error diagnosis and explained answersDemonstrate
Audit a metric and defend a reliable business recommendation.
Queries executed and results compared automatically · Final assessment and business projectYour progress
Start the path
Detailed syllabus
10 structured lessons
Move from a business question to controlled, explainable and actionable SQL indicators.
Detailed syllabus 1Make data and indicators reliable5 structured lessons+
- 01Choose the level of detail and prevent double countingSecure a KPI before a join between multiple 1:N relationships. · Related practice: Avoid double counting after several joins
- 02Building cohorts and measuring retentionTrack the same counts over time with a stable denominator. · Related practice: Building a cohort and measuring retention
- 03Segment a portfolio with RFMTransform recency, frequency and value into explainable business rules. · Related practice: Build a complete RFM segmentation
- 04Measure a funnel without inflating conversionsMove from a noisy event log to a controlled funnel. · Related practice: Measure a funnel without counting duplicate events
- 05Master windows, partitions and framesDistinguish between PARTITION BY, ORDER BY, ROWS and RANGE to avoid ambiguous indicators. · Related practice: Analyze a time series with missing months
Detailed syllabus 2Analyse, compare and defend a conclusion5 structured lessons+
- 06Deduplicate with ROW_NUMBER without losing the truthChoose a winning line with a complete and auditable rule. · Related practice: Deduplicate with a deterministic rule
- 07Analyze margin and contribution without confusing revenue and profitLink sales, discounts and costs at the right level of detail. · Related practice: Analyze the margin and real contribution
- 08Build a KPI with its quality controlsDeliver a figure accompanied by its scope and consistency tests. · Related practice: Audit a KPI before publication
- 09Auditing an imperfect datasetProfile NULL, formats, duplicates, domains and breaks before analysis. · Related practice: Clean an imperfect dataset before analysis
- 10Conduct a standalone SQL analysisMove from an ambiguous question to a repeatable recommendation. · Related practice: Conduct an autonomous seller analysis
End-of-path practice
Put the skills from this path into practice.
Revisit an exercise at this level, explain the result and check an edge case. This practice is included in your path; it is not an independent certification.
Avoid double counting after several joins